Certain devices of the this type have been proposed by the patents France Nos. 1 382 916 and 2 494 988, and their equivalents U.S. Pat. No. 3,385,937 and European Patent Publication No. 54,450, respectively.
It is a particular object of the invention to render devices of the type concerned such that the sounds transposed by them into the bass range of the spectrum comprise information relating not only to the pitch and to the intensity of the sound signals to be perceived, but also the timbres or laryngeal components of these signals.
Accordingly, auditive prosthetic devices of the above type comprise still, in manner known in itself, transducer means to transform the acoustic signals to be processed into electrical signals, means for sampling at least one particular slice x.sub.i of order i of each of these electrical signals, with the slice being the frequency components comprised in a pre-determined range of high frequencies, that is to say corresponding to high-pitched sounds, frequencies comprised between f.sub.i min and f.sub.i max, i being an integer comprised between 1 and 3 and increasing in the same direction as the frequencies of the slices concerned, means for elaborating from each of these slices x.sub.i a compensation signal S.sub.i comprised within a range of low frequencies, that is to say corresponding to bass sounds, frequencies comprised between F.sub.i min and F.sub.i max, the bass ranges corresponding to the various compensation sounds S.sub.i, being distinct from one another, means for adding together the different compensation signals S.sub.i thus elaborated and transducer means to transform the sum-signal so-obtained to a signal applicable to the deaf person to be treated so as to be perceptible by him.
According to the invention, the hearing aid devices so-defined are characterized in that the means for elaborating each compensation signal S.sub.i comprise means for multiplying the corresponding signal slice x.sub.i by a signal x'.sub.i which is either identical with this signal slice x.sub.i or deduced from the latter by amplitude equalization, so as to form a product-signal X.sub.i, and means for sampling, in this product-signal X.sub.i, by means of a band-pass filter, the portion of said product-signal of which the low frequency is comprised within the range of frequencies F.sub.i min to F.sub.i max, which portion constitutes the desired compensation signal S.sub.i.
Applicants have in fact discovered that the product-signals X.sub.i concerned comprise such portions at low frequency by reason of properties ascribable to the products of trigonometric lines.
It happens in fact that each of the signals x.sub.i can be decomposed into a Fourier series, that is to say into a multitude of sinusoidal components of the type EQU a cos 2.pi.ft+b sin 2.pi.ft (1)
in which formula a and b denote constants and f the oscillation frequency of the component concerned, said frequency being comprised between the two limits f.sub.i min and f.sub.i max.
If one considers, in addition to the component which has just been defined, a second component comprised by the same signal x.sub.i and defined by the formula EQU a' cos 2.pi.f't+b' sin 2.pi.f't (2)
in which a' and b' denote also two constants and f', the frequency of this second component comprised like f between the two limits f.sub.i min and f.sub.i max, and if one multiplies these two components by one another, the product shows not only terms in cos 2.pi. (f+f')t and sin 2.pi. (f+f')t but also terms in cos 2.pi. (f-f')t and sin 2.pi. (f-f')t.
More generally, if the signal x.sub.i is multiplied by itself, the product signal X.sub.i shows sinusoidal components oscillating at almost all the frequencies of the type f-f' corresponding to differences of any two frequencies comprised within the range of f.sub.i min to f.sub.i max.
Among the innumerable differences in frequency so-defined, a certain number necessarily occur within the ranges of low frequencies desired for the compensating bass sounds, and it is this which enables such bass sounds to be obtained, said sounds respecting wholly the temporal forms of the high-pitched sounds to be represented.
This respect for temporal forms enables the transposition towards the low notes not only of the high notes and intensities of the high-pitched slices sampled, but also their timbres, which increases considerably the intelligibility of the processed sounds.
In advantageous embodiments, recourse is had in addition to one and/or other of the following features:
the number n of the slices of signal x.sub.i at high frequency which are sampled in the electrical signals to be processed is equal to two, the first slice being comprised between 1500 and 3500 Hz and the second, between 5000 and 7000 Hz,
in an arrangement according to the preceding paragraph, the ranges of low frequencies selected for the compensation sounds S.sub.1 and S.sub.2 are respectively from 800 to 1000 Hz and from 1050 to 1200 Hz,
in an arrangement according to the the next-to-last paragraph, the ranges of low frequencies selected for the compensation sounds S.sub.1 and S.sub.2 are respectively from 800 to 1200 Hz and from 1250 to 1900 Hz.
The invention comprises, apart from these main features, certain other features which are preferably used at the same time and which will be more explicitly considered below.